[1] opecv: opecv中的坐标系_opencv坐标系

opencv中的坐标系

【1】opencv中的坐标系

opencv中，以图像image建立坐标系：
原点(0， 0)：图像image左上角像素点；
水平方向为x轴（即图像的宽、列column），竖直方向为y轴（即图像高、行row）。

cv2.read(‘图像’).shape --> 返回：(height, width, channel)，先高后宽，即(y, x, channel)，先y后x；
但是在图像操作输入坐标点时，输入的是(x, y)，先x后y，即(w, h)，如：
img_resize = cv2.resize(img, (w, h))
cv2.rectangle(img0, (x0, y0), (x1, y1), (0, 0, 255), 2)
cv2.rectangle(img0, (0, 0), (50, 300), (0, 0, 255), 2)

opecv中像素点的坐标：
----------------------------------->x
∣
∣ (x1,y1)−−−−−−------
∣ ∣ |
∣ ∣ |
∣ ∣ |
∣ ∣−−−−−−−−------(x2,y2)
|
∨
y

【2】读取image后为numpy ndarray类型，np与opencv中的坐标点(x, y)是转置T关系
如，把opencv坐标系中的ROI(0,0) -> (50,300)区域像素改为黄色：
img[0:300, 0:50] = [0, 255, 255] # 先y后x

【3】opencv坐标系coordinate与画图draw测试代码

import os
import cv2


def draw_test(img):
    '''
    draw rectangle: in opencv coordinate rectangle area: ROI(0,0) -> (50,300)
    (x1, y1) (x2, y2) is in opencv coordinate rectangle
    '''
    img0 = img.copy()
    cv2.rectangle(img0, (0, 0), (50, 300), (0, 0, 255), 2)
    # cv2.rectangle(img0, (0, 0), (50, 300), (0, 0, 255), -1)

    cv2.imshow("imgdraw", img0)
    cv2.waitKey(1000)


def resize_test(img):
    '''
    input: img
    output: height zoom out to 480, weight zoom out according ratio in opencv coordinate
    '''
    h, w = img.shape[0:2]
    hzoom = 480
    wzoom = int(w * hzoom / h)

    imgzoom = cv2.resize(img, (wzoom, hzoom))

    cv2.imshow("img zoom out", imgzoom)
    cv2.waitKey(1000)


def roi_test(img):
    '''
    input: img
    '''
    img0 = img.copy()
    h,w = img0.shape[0:2]

    #in opencv coordinate rectangle area: ROI(0,0) -> (50,300), change to yello
    img0[0:300, 0:50] = [0, 255, 255]  # numpy`s coordinate (x,y) is Transposed relation shape with opencv
    # get pix value of (25, 25), (400, 240)
    print("(25, 25) pix value:", img0[25, 25])
    print("(400, 240) pix value:", img0[400, 240])

    cv2.imshow("imgROI", img0)
    cv2.waitKey(1000)


def coordinate_test(img):
    '''
    input: img
    '''
    height, width, ch = img.shape[:]  # (row,col,ch) (height,width,ch)
    print("img type:%s, img size:%s" % (type(img), img.size))  # img.size : all pixels count
    print("image shape:%s, shape type:%s" % (img.shape, type(img.shape)))
    print("height:%s, width:%d, ch:%d" % (height, width, ch))

    cv2.imshow("imgcoordinate", img)
    cv2.waitKey(1000)

if __name__ == "__main__":
    # print(os.getcwd())
    img = cv2.imread("zta.png")

    coordinate_test(img)
    roi_test(img)
    resize_test(img)
    draw_test(img)


#img type:<class 'numpy.ndarray'>, img size:966036
#image shape:(722, 446, 3), shape type:<class 'tuple'>
#height:722, width:446, ch:3
#(25, 25) pix value: [ 0 255 255]
#(400, 240) pix value: [ 45 35 204]
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76